Meta: Muse Spark 1.1 now available on OpenRouter (1,049k context, $1.25/M in, $4.25/M out)
TL;DR
Meta's Muse Spark 1.1 multimodal reasoning model is now available on OpenRouter with 1,049k context at $1.25/M input and $4.25/M output.
What changed
Meta made Muse Spark 1.1 available through OpenRouter with a 1,049k context window and pricing at $1.25 per million input tokens and $4.25 per million output tokens. Developers can now route multimodal inputs including text, images, video, audio, and PDFs to the model for agentic tasks. Basic Users and Vibe Builders gain direct access without separate setup.
Why it matters
Developers working on agentic tasks benefit from the 1,049k context that handles long PDF documents and mixed media in one pass, a scale that exceeds many current options from competitors like standard Llama deployments. Vibe Builders can explore detailed multimodal reasoning flows while Basic Users test the model through simple OpenRouter calls at the stated rates.
What to watch for
Compare results against Claude 3.5 Sonnet on the same OpenRouter routes for agentic workflows. Developers should run a sample PDF plus image query through the model and review output length against the 1,049k limit listed on the provider page.
Who this matters for
- Vibe Builders: Route your mixed PDFs and images through OpenRouter to test multimodal reasoning flows.
Harsh’s take
Meta's Muse Spark 1.1 on OpenRouter offers a highly competitive price point for a million-token context window. At $1.25 in and $4.25 out, it undercuts several proprietary models while handling text, audio, video, and PDFs in a single call. Operators should immediately benchmark this against Claude 3.5 Sonnet for multi-document agentic tasks. The massive context window is only useful if the model actually maintains high retrieval accuracy across the entire span without hallucinating details.
by Harsh Desai
About Modal
View the full Modal page →All Modal updatesGo deeper
More from Modal
- Model ReleaseKimi releases K3 open model with 2.8T parameters and 1M context
Kimi launches K3, a multimodal open-weight model with 2.8 trillion parameters and 1M token context. It nears GPT-5.6 Sol and Claude Fable 5 in benchmarks, with full weights due by July 27.